1. Field of the Invention
This invention relates to an improvement to eliminate the need for slip rings in calibration system utilizing a signal processor that receives the outputs of one or more motion sensing devices mounted on a rotary platform. The devices are subjected to known conditions of rotary motion and the outputs are compared with the known conditions to determine any errors therebetween to produce correction factors that are applied to the outputs to more accurately indicate the known conditions. In the preferred embodiment, the invention is used with calibration equipment for inertial measurement units (IMU""s) mounted on a rate table and, more particularly, to a self contained system that avoids the use of slip ring connections for transferring signals from the IMU""s to the signal processor.
2. Description of the Prior Art
Heretofore, the calibration of IMU""s has required the mounting of the IMU under test on a multi-axis rate table, subjecting the IMU to predetermined known motions (e.g. rates and accelerations) in the various axes (e.g. three perpendicular axes such as pitch, roll and yaw), and at various known temperatures and other environmental conditions, determining the response of the IMU to these input environmental condition values and transmitting the output signals indicative of such values through a slip ring assembly to a remote computer or processor which includes test software that operates to compare the IMU outputs with the predetermined known values and to determine any errors so as to produce calibration coefficients which can be used to correct the errors in the outputs. The calibration coefficients are stored in a memory, internal to the IMU, and which are then used with calibration software in the IMU to add or subtract the correction values from future IMU outputs, at the various rates and temperatures, etc.
While it is not necessary for an understanding of the invention, it is believed helpful to consider the operation of the prior art in establishing the correction values. Three major causes of inaccuracy in an IMU are scale factor, bias and misalignment. These three factors may also vary with temperature so temperature drift may be considered a fourth factor. There are others, but these tend to be the most important to be corrected by the calibration system and will be used here for simplicity. Again, for simplicity, assume that a single axis sensor, such as an accelerometer, is mounted on a base that will be used in actual practice and assume that this base is mounted on a rate table and in an oven. If the base is mounted so that the sensing axis of the accelerometer is parallel to the gravitational axis, the IMU should produce a signal indicating xe2x80x9c+1 gxe2x80x9d when upright and xe2x80x9cxe2x88x921 gxe2x80x9d when inverted, without any motion of the rate table. (Outputs greater than 1 g are produced by rotation of a table such as would be experienced on a centrifuge). If there is a scale factor error of, say, +1.0%, then at +1 g, the output would read +1.01 g and at a xe2x88x921 g, the output would read xe2x88x921.01 g. Accordingly, to correct this error, one would multiply the output by 1.00/1.01=0.99. If there was a bias error of say +0.2 g, then at +1 g, the output would be 1.2 g and at xe2x88x921 g the output would be xe2x88x920.8 g. To correct this error, the value (1.2+(xe2x88x920.8))/2=0.4/2=0.2 g would be subtracted from the output. Misalignment errors are normally determined when the sensing axis is perpendicular to the gravitational field where a perfect alignment would produce an output of 0.0 g. The correction factor for misalignment is determined by observing the acceleration error from the accelerometer as the accelerometer is rotated about its input mounting axis placed in the 0 g orientation. The misalignment correction factor is then calculated by using a small percentage of the other two acceleration axes outputs of an orthogonal system.
All of the tests will then be re-conducted at a plurality of various temperatures so that the memory will contain the proper correction coefficients for all temperatures to be encountered in actual use. It should be noted that in addition to temperature, there are other environmental conditions, such as humidity, altitude, and vibration that may affect the outputs.
In all cases, all of the output signals must be sent through the slip ring assembly to the processor where the correction coefficients are determined and then all of the correction coefficients must be transmitted from the processor back through the slip ring assembly to the memory in the IMU.
Now working at an IMU level with three axis accelerometers, then the simplified tests above will have to be repeated along all three of the axis of interest but, although more complicated, the result will still be a lookup table in the IMU which will correct for the major errors in all three axes and at all temperatures to be encountered. It should be understood that the tests may use more advanced algorithms than the simplified equations above, to detect the errors along the desired axes. For example, by aligning the IMU along axes that are skew to the three major axes, a new vector output may be obtained. Then the new vectors may be mathematically resolved to provide the desired information along the three major axes without as many repetitions of the application of various g forces to the IMU.
In any event, and regardless of the processor algorithm used, the requirement for using a slip ring assembly to read the IMU output values presents a considerable problem since such assemblies are very expensive particularly, when it is desired to calibrate a large plurality of IMU""s (e.g. 32 units) at the same time.
The present invention eliminates the need for slip-ring assemblies to obtain the signals from the IMU""s for presentation to the external processor by building the IMU""s with internal processors capable of using the test software to determine the calibration coefficients and the calibration software to modify the IMU outputs for customer use. This also eliminates the need for external remote processors and simplifies the calibration process at a greatly reduced cost.